AI screen recording for QA: one recording, one reproducible GitHub issue

Here is how a QA tester turns one recording into a reproducible GitHub issue, with Vidmatic. QA spends as long writing defects up as finding them. Reproduction steps get lost, screenshots go stale, and tickets bounce back for missing detail. Set it up once. In settings, connect GitHub, pick a default repository, and give each repository its own default labels and assignees. Start a recording from the Chrome extension with auto-create GitHub issue turned on. Say the expected result out loud, just before the step that fails. Circle the wrong value with the pen, or write the right one by hand. Each mark is saved as an image at the moment you confirm it. When you stop, Vidmatic transcribes your voice and opens one issue in your default repository, with its labels, its assignees, and a share link to the run. Anyone who can read the issue can watch the run, signed in or not. A developer can hand the same recording to Claude Code, with every word and mark. When the developer's agent records a before and an after, Vidmatic plays the pair back to back, so you can check the fix without setting up the build yourself. No extension? New recording in the app records any screen or window, a desktop build too. There is no pen there, so name the element out loud.